PREDICTIVE ANALYTICS LOS ANGELES COUNTY CSSD DR. STEVEN GOLIGHTLY DIRECTOR LOS ANGLES COUNTY CHILD SUPPORT SERVICES
Los Angeles County LA County is 24% of the State s child support caseload. Over 1500 employees Caseload sizes averages 800 per case caseworker
Data Analytics Analytics is the application of computer technology, operational research, and statistics to solve problems in business and industry. For Los Angeles: Program Support Division Programmers and Analysts who research and create statistical reports for management review. Includes a training unit that puts into practice decisions made by management and then evaluates the effect that policy or training has had on the department CSTATs Meeting Attended by all senior management where departmental performance is reviewed, discussed, and decisions made targeting areas for emphasis or improvement Data Sharing Down to the line level, data is provided for each worker to research and improve their own caseload, performance, and information sharing of best practices. SPSS Software program used for analysis and statistics
Predictive Analytics Encompasses a variety of techniques that analyze current and historical facts to make predictions about future, or otherwise unknown, events. Statistics Modeling Data mining
Model Development Procurement of software Staffing concerns Training on software
LA County SPSS Procurement Process Model Development LA County CSSD chose IBM s SPSS Modeler Professional as our predictive analytic tool since we are already working with SPSS Statistics. Client license vs. Desktop license o Modeler could be used by a number of users from their own computers. Cost o Purchased one license, which will require users to schedule when the software can be used.
Model Development Staffing Concerns o Skill set needed for analytics o Internal staff (building up skills in analytics) o Hiring new staff with this kind of background o Unions Training of Software
Predictive Analytics Goal: By June 30, 2014, develop an initial analytical model which can be used to predict the probability that payments will be received in a child support case. Score all active enforcement cases based on those predictions and distribute them in a format that will allow Child Support officers to prioritize their work. Goals
Child Support STATS (CSTATS) Case Segmentation Survival Study Project with IBM and UC Berkley Program Efforts
Survivor Study Research Project Goal How important is the receipt of child support to formerly assisted families in remaining selfsufficient and not reliant upon public assistance? Does regular payments or the amount of child support paid affect a formerly assisted custodial parent s ability to remain off of CalWORKs? Establish an initial baseline data set for future studies to build upon, both on this topic and new topics, by highlighting questions and data elements that may arise.
Survivor Population Overview Given all the independent variables previously discussed, the most common welfare recipient would be: CP Role Mother (75%) Gender Female (90%) Ethnicity Hispanic (38%), White (21%), black (9%) Primary Language English (79%, includes English as second language) CPs Average Age 36 years old NCPs Average Age 37 years old Number of Children 12 Children s average Age 11 years old
Survival Results Interval Start Time Number Entering Interval Number Withdrawi ng during Interval Number Exposed to Risk Number of Terminal Events Proportion Terminatin Proportion g Surviving Cumulative Proportion Surviving at End of Interval Probability Density Std. Error of Probability Density Hazard Rate Std. Error of Hazard Rate 0 22515 0 22515.000 0.00 1.00 1.00.000.000.00.00 1 22515 0 22515.000 12327.55.45.45.548.003.75.01 2 10188 0 10188.000 5569.55.45.21.247.003.75.01 3 4619 0 4619.000 1256.27.73.15.056.002.31.01 4 3363 0 3363.000 583.17.83.12.026.001.19.01 5 2780 0 2780.000 450.16.84.10.020.001.18.01 6 2330 0 2330.000 248.11.89.09.011.001.11.01 7 2082 0 2082.000 239.11.89.08.011.001.12.01 8 1843 0 1843.000 153.08.92.08.007.001.09.01 9 1690 0 1690.000 106.06.94.07.005.000.06.01 10 1584 0 1584.000 98.06.94.07.004.000.06.01 11 1486 0 1486.000 46.03.97.06.002.000.03.00 12 1440 0 1440.000 58.04.96.06.003.000.04.01 13 1382 0 1382.000 39.03.97.06.002.000.03.00 14 1343 0 1343.000 44.03.97.06.002.000.03.01 15 1299 0 1299.000 28.02.98.06.001.000.02.00 16 1271 0 1271.000 20.02.98.06.001.000.02.00 17 1251 0 1251.000 40.03.97.05.002.000.03.01 18 1211 0 1211.000 38.03.97.05.002.000.03.01 19 1173 0 1173.000 31.03.97.05.001.000.03.00 20 1142 0 1142.000 28.02.98.05.001.000.02.00 21 1114 0 1114.000 32.03.97.05.001.000.03.01 22 1082 0 1082.000 18.02.98.05.001.000.02.00 23 1064 0 1064.000 22.02.98.05.001.000.02.00 24 1042 0 1042.000 0.00 1.00.05.000.000.00.00 25 1042 0 1042.000 1042 1.00.00.00.000.000.00.00 a. The median survival time is 1.91
Challenges Staffing Civil Service Rules Unions Case ownership model
CSSD 2017 What predictive analytics will look like in like in 4 years.
THANK YOU FOR YOUR TIME Dr. Steven Golightly Steven_Golightly@cssd.lacounty.gov (323) 8893400